1. Technical Field
The present invention relates generally to gain compensation in an amplifier circuit and, more particularly, to an equalizer having a frequency dependent gain for multi-octave passband equalization.
2. Discussion
May communications systems operate over a wide range of radio frequency (RF) input frequencies. Where the maximum frequency is at least double the minimum frequency, the communication system is considered multi-active and presents special considerations. For example, a L-band digital receiver can receive frequencies over the range of 600 megahertz (MHz) to 1,800 MHz. Such a typical digital receiver includes an amplifier for sufficiently amplifying an input signal for processing. Similarly, many RF transmitters include an amplifier for sufficiently amplifying the signal prior to output by the transmitter.
In order to achieve the optimal noise and linearity required by receivers and transmitters, two wideband commercial amplifiers are typically cascaded in a chain. The first amplifier in the chain preferably has a low noise characteristic and moderate linearity. The second amplifier in the chain preferably provides a better noise characteristic and improved linearity. An impedance matching network is inserted between each amplifier stage to improve the input and output voltage standing wave ratio (VSWR).
Such amplifier chains typically include a pair of amplifiers, each providing a specific, predetermined amplification. For example, the L-band digital receiver typically includes a front end amplifier chain which provides high linearity, low noise, and a nominal gain of 21 decibels (db) across a band of approximately 500 MHz to 1,900 MHz. To achieve optimal noise and linearity, two wideband commercial amplifiers are selected. The first stage has a low noise and moderate linearity. The second stage has good noise performance and high linearity. The amplifier chain is typically implemented using a low cost, compact, high performance, wide band monolithic microwave integrated circuit (MMIC).
Each MMIC amplifier in the amplifier chain of the L-band digital receiver has a modest gain slope across the input or passband range of 600 MHz to 1,800 MHz. Each amplifier typically provides a falloff or gain of approximately 1.5 to 2.0 db over approximately 500 MHz to 2 gigahertz (GHz). The falloff in gain results in a negative gain slope over an increasing frequency. Such a gain slope is characteristic of multi-octave wideband package MMIC amplifiers or gain blocks. Because each amplifier or gain block is arranged for operation with a matched, 50 ohm system, passband equalization using lossless reactive elements was not possible.
The above-described amplifier chain may be modeled with a standard 4 db attenuator pad placed between each amplifier stage. Such modeling results in a well matched network with gains nominally sloping from 24 db to 21 db over the frequency range of interest. The gain of the amplifier network is preferably set to establish a minimum noise value at the high end of the frequency range where the gain is the lowest. That is, the high end of the frequency range preferably has a noise figure or additive noise of approximately 3.2 db. However, the linearity of the network at the high end of the frequency passband is nominally 13 db relative to an input power of approximately 1 milliwatt. At the low end of the frequency range where the gain is highest, the linearity is significantly degraded to a nominal level of approximately 10.8 decibels relative to an input power of one milliwatt because the second stage amplifiers are overdriven by first stage amplifier.
In the above-described networks, there is a tradeoff between the gain, the noise, and the resistance to distortion of the amplifier network. In a typical MMIC amplifier used in a network for transmitters and receivers operates on frequencies where the high frequency is at least double the lowest frequency, the gain of the amplifier decreases as the frequency decreases. When the gain decreases, the noise of the amplifier increases. Further, in multi-stage amplification networks, even though the gain of the first stage must be set at some minimum, the gain must be limited in order to maintain linearity of the second stage amplifier. At the lower end of the frequency range, the relatively high gain and low frequency significantly impact the linearity of the second stage amplifier. Further, specifications for each amplification stage of the network typically require a generally even balance of gain across the entire frequency band at each stage.
Thus, it is desirable to provide a passband equalization network which preserves linearity and noise requirements while simultaneously maintaining proper impedance matching. In the example described above, is desirable to flatten the gain passband of the first stage amplifier so as to overdrive the second stage amplifier. The slope of the entire network would have to be fairly flat and linear over the operating frequency range. Further, such a network must be impedance matched with maximum return loss in order to avoid inducing passband ripple caused by reflected RF energy.
The present invention is directed to an amplifier network including a first amplifier receiving an input signal and generating a first output signal. An equalizer receives the output signal and attenuates the first output signal in accordance with the frequency of the input signal. The equalizer generates an equalized signal. A second amplifier receives the equalized signal and generates an output signal for the network. The equalizer includes a PI network and a TEE network interconnected in a symmetric configuration.
For a more complete understanding of the invention, its objects and advantages, reference should be made to the following specification and to the accompanying drawings.